A Custom 3D-Printed Mold to Make MRI-Guided Heat Therapy for Prostate Cancer More Precise

Table of Contents

• Key Points
• Why This Research Matters
• Background: Radiation, Heat, and the Need for Perfect Positioning
• What the Researchers Wanted to Achieve
• Study Methods: How the Research Was Conducted
• Building the Custom Devices: Scanning, Design, and 3D Printing
• How Positioning Accuracy Was Measured
• Key Findings: Reproducibility at the Radiation Machine (LINAC)
• Key Findings: Reproducibility on MRI
• The Crucial Comparison: With Device vs. Without Device
• Ultrasound Targeting: Can the Device Aim the Heat Correctly?
• Clinical Implications: What This Means for Patients
• Limitations: What This Study Could Not Prove
• Recommendations: What Patients Should Know
• Frequently Asked Questions
• Source Information

Key Points

• A research prototype 3D-printed pelvic mold reduced positioning errors by 4–6 times versus tattoo-based alignment in two healthy volunteers.
• Most positioning errors with the mold were below half a millimeter; the largest occurred along the head-to-feet direction.
• The mold held an ultrasound probe so that target vessels appeared within 2 mm of their reference position across sessions.
• The device was tested on two healthy volunteers only; no actual prostate cancer patients were included, and no heat therapy was given.
• Printing each mold took 65–77 hours, but the mold enabled reproducible positioning without tattoos or lasers.

Why This Research Matters

Prostate cancer is the most common cancer in men. Many men who have had their prostate removed surgically (radical prostatectomy) later experience a "biochemical relapse" — a rise in prostate-specific antigen (PSA) blood levels indicating that cancer cells may still be present.

For these patients, external beam radiotherapy (EBRT) — precisely targeted radiation from outside the body — is the standard curative treatment. But long-term disease control remains limited. Roughly 50% of patients (about 1 in 2) experience disease progression within 5 years of salvage radiation.

Doctors have tried giving higher radiation doses to improve results. However, when radiation doses exceed 70 Gray (Gy), long-term side effects may increase. This is a serious concern, because radiation must kill cancer cells while protecting nearby healthy organs such as the bladder and rectum.

Background: Radiation, Heat, and the Need for Perfect Positioning

Hyperthermia (HT) — a treatment that gently heats tissue to between 41°C and 43°C (106–109°F) — is a clinically accepted way to make radiation more effective. Heat damages cancer cells and makes them more vulnerable to radiation, potentially improving cell killing. In theory, adding heat could allow doctors to achieve better results without escalating the radiation dose, which would spare patients from added side effects.

Delivering deep heat is technically challenging. A technology called magnetic resonance-guided focused ultrasound (MRgFUS) uses sound waves that concentrate energy sharply inside deep tissues, guided in real time by MRI imaging. Over the past decade, studies have shown it can be used for tissue ablation (destroying tumors with heat) and for hyperthermia in various tumor sites.

Both radiation therapy and ultrasound heat therapy demand exceptional precision. The patient must lie in exactly the same position during every session — sometimes for weeks of daily treatments. Radiation oncology departments currently achieve this with tattoos: small permanent ink dots on the skin that are lined up with laser beams on the treatment machine. For prostate cancer, correct daily repositioning has been linked to better long-term disease control and fewer radiation-induced side effects.

But tattoos have downsides. They can be a constant reminder of disease, and tattoo removal is often not covered by health insurance. More importantly for this research, tattoos alone cannot hold a patient still enough, and they certainly cannot hold a heavy ultrasound applicator in place. Standard commercial immobilization devices — such as Vac-Lok™ cushions (CIVCO, IA, USA), Alpha Cradle® (Smithers Medical Products, OH, USA), thermoplastic masks, and knee and ankle supports — also cannot embed ultrasound applicators.

This is where 3D printing comes in. It can produce personalized, rigid devices that immobilize the patient and integrate dedicated holders for hyperthermia applicators. A rigid 3D-printed shell also supports the weight of the MRI coil, preventing body deformation. This improves MRI quality and MR thermometry (temperature measurement during MRI) by reducing the distance between the body and the coil and by reducing motion.

What the Researchers Wanted to Achieve

This study was part of an ongoing clinical project exploring the combination of MRgFUS-based hyperthermia and salvage radiation therapy in prostate cancer patients with biochemical relapse after radical prostatectomy. The team's goal was to develop and test a personalized 3D-printed immobilization device that could be used for both treatments and that had a built-in holder for an ultrasound applicator.

Specifically, they wanted to test whether such a device could achieve:

• Fast, reproducible patient repositioning during repeated radiation and hyperthermia sessions, and
• Reliable ultrasound targeting of a predefined region, so a future therapeutic ultrasound beam could aim at the same tissue every time.

Notably, all positioning tests were done without using skin marks, tattoos, lasers, or table shifts to align the volunteers — the physical shape of the mold alone was supposed to do the job. Two healthy volunteers with very different body types (body mass index, or BMI, of 22 and 31) took part in the study, to prove the system works for different body shapes.

Study Methods: How the Research Was Conducted

Each volunteer first lay on their back (the supine position, standard for pelvic radiation treatment) on a dedicated table. A conventional ultrasound imaging probe (CHISON Medical Technologies, Sonobook 6, P2-V probe, Jiangsu, China) was held still by an articulated arm while an expert radiologist captured reference ultrasound images.

Abdominal pulsed color Doppler imaging focused on the right iliac artery — a major blood vessel in the pelvis — was used as an internal organ landmark. This meant that the researchers could later check whether the ultrasound probe embedded in the mold saw the same vessel in the same place.

A portable, metrology-grade 3D surface scanner (HandySCAN™ 300, Creaform, QC, Canada) was used to create a digital model of each volunteer's pelvic region and the ultrasound probe. Its specifications are highly precise:

• Optical resolution: 0.025 mm (25 millionths of a meter)
• Mesh resolution: 0.1 mm
• Sampling rate: 205 kHz
• Light source: 3 laser diodes
• Stand-off distance: 300 mm
• Depth of field: 250 mm

The resulting surface image was cleaned and extruded (built up) to a thickness of 5 mm using VxElements™ software (Creaform, Canada).

Building the Custom Devices: Scanning, Design, and 3D Printing

The researchers created a 3D model of the prototype device that recorded the ultrasound probe's exact angle and position. The device was then printed using ABS (acrylonitrile–butadiene–styrene) plastic — a material chosen because it is both MRI-compatible and resistant to radiation. The printer was an A2v4 machine (3NTR, Oleggio, Italy).

The finished devices were each printed in two parts (a top and a bottom shell). Production figures show the substantial effort involved:

• Volunteer 1 (BMI 22): 65 hours of printing; each piece measured approximately 35 cm × 33 cm × 12 cm; total device weight 1.3 kg.
• Volunteer 2 (BMI 31): 77 hours of printing; each piece measured approximately 51.5 cm × 35 cm × 16 cm; total device weight 1.7 kg.

Quality assurance was rigorous. Every printed part was scanned with the same high-accuracy HandySCAN 300 scanner used for the original model. The scans were aligned and compared with the digital models using VxElements software to produce a surface deviation map — essentially a color-coded map showing how closely the real printed object matched the perfect digital design.

The deviation distribution was almost Gaussian (bell-shaped), with a mean of zero for both volunteers. For the first volunteer, the standard deviation was 0.3 mm, with maximum deviations under 0.7 mm. For the second, the standard deviation was 0.5 mm, with maximum deviations under 1.0 mm. In plain terms, the 3D printer reproduced the digital designs with sub-millimeter accuracy.

The device's internal quality was checked with a computed tomography (CT) scan (Philips Big Bore, Utrecht, the Netherlands) using 1 mm slice thickness. The CT density of the plastic was uniform in all volunteers, with a mean of -350 Hounsfield units (HU) and a standard deviation of ±125. A minimum threshold of -650 HU was chosen to verify that the shell thickness matched its physical thickness. This scan would also allow doctors to create a "dummy" reference surface for the optical monitoring system used during treatment.

How Positioning Accuracy Was Measured

The researchers tested how precisely each volunteer could be repositioned in the mold, using two measurement systems at two different machines: an optical surface monitoring (OSM) system at the linear accelerator (LINAC, the radiation machine) and a 3 Tesla MRI scanner (Magnetom Prisma Fit, Siemens, Erlangen, Germany).

At the LINAC (a Truebeam system, Varian Medical Systems, CA, USA), the OSM system used was AlignRT™ (VisionRT, London, UK). The device was attached to the LINAC table using a standard Vac-Lok cushion, which was molded to fit the bottom shell and served as a reference frame. On the MRI table, a more compact cushion was needed because the standard Vac-Lok system was too large and collided with the MRI ring.

Three MRI-compatible aqueous gel markers were fixed to the bottom shell segment. These allowed the researchers to check for two kinds of movement: movement of the volunteer relative to the shell, and movement of the shell relative to the MRI table.

In each measurement session (at the LINAC or MRI), the volunteer was asked to get into the bottom shell, and the top part was added. The volunteer then got out and repositioned himself between each of 6 or 7 repeated measurement sequences. No lasers or skin marks were used — the table was simply placed at the same vertical, longitudinal, and lateral coordinates each time. The LINAC's OSM system — which has sub-millimetric resolution — recorded three position offsets (vertical, horizontal, and longitudinal) and three rotation angles (yaw, roll, and pitch).

For the MRI sessions, position differences between a reference 3D image and the other acquisitions were calculated on the Eclipse™ treatment planning system (Varian Medical Systems, Palo Alto, CA, USA). An automatic algorithm co-registered the images based on bony anatomy (the patient's bones, which do not move or deform). The fourth image of each session was used as the reference to account for any pelvic anatomical changes (such as bladder or rectal filling) during the approximately 40-minute acquisition of seven images.

The MRI sequence parameters tell the story of how detailed these images were:

• T1-weighted turbo spin echo sequence
• Repetition time (TR) = 784 ms
• Echo time (TE) = 12 ms
• Bandwidth = 170 Hz/pixel
• Turbo spin echo factor = 3
• Refocusing pulse = 180°
• Physical resolution = 1.56 mm × 1.17 mm × 3 mm
• GRAPPA acceleration = 2
• Field of view = 300 mm
• Phase oversampling = 80%
• Number of signals averaged = 2
• Acquisition time = 185 seconds per image
• Coil combination: spine and body matrix with 30 elements

An expert radiologist also measured the distance between the MRI markers on the shell and the lower anterior corner of a specific sacral vertebra (the S4 vertebra for volunteer 1 and the S2 vertebra for volunteer 2) on sagittal images. This measured whether the patient's bones stayed in a fixed relationship to the rigid shell itself.

Statistical comparison of the 3D error vectors between the OSM data and MRI data used the Mann–Whitney test, with p-values below 0.05 considered statistically significant.

Key Findings: Reproducibility at the Radiation Machine (LINAC)

When measured by the OSM system at the LINAC, positioning errors in the vertical (up-down) and lateral (side-to-side) directions were minimal for both volunteers: median values were smaller than 1 mm, with a range of about 2 mm. Rotations around three axes were also small, with median values under one-third of a degree and a range of about 1°.

The longitudinal direction (along the length of the patient's body, head-to-feet) proved more challenging — a finding consistent with earlier research. For volunteer 1 (BMI 22), longitudinal deviations were larger, with a median of 4.4 mm and a range of 18.8 mm. The 3D vector (the combined error in all directions) had a median of 4.5 mm for this volunteer.

Volunteer 2 (BMI 31) actually did much better in this direction, with longitudinal deviations smaller than 0.7 mm and a range of just 1.2 mm, thanks to better fixation against the couch. His overall 3D vector was a median of 0.8 mm. However, his rotation values were larger — a median of almost 1° with a range of 1.5° — likely because this volunteer lost weight between the 3D scanning session and the OSM measurements (his BMI dropped from 31.4 to 29.8), meaning the mold fit slightly less snugly.

Key Findings: Reproducibility on MRI

Using the 3D-printed device, reproducibility based on bony anatomy was excellent for both volunteers on MRI. Lateral and vertical deviations were very small:

• Volunteer 1 (BMI 22): median under 0.5 mm, range about 2.5 mm
• Volunteer 2 (BMI 31): median 0.5 mm, range 3.8 mm

As at the LINAC, longitudinal deviations were the largest. Both volunteers had a median longitudinal deviation of 1.4 mm, with ranges of 8.5 mm (volunteer 1) and 7.5 mm (volunteer 2). Rotations were small: near zero with a range of about 1° for volunteer 1, and near 0.5° with a range of about 2° for volunteer 2.

The overall 3D displacement vectors on MRI were a median of 1.5 mm for volunteer 1 and 2.3 mm for volunteer 2.

The measurements of the bone-to-marker distances confirmed that the patient's bony anatomy stayed reproducible relative to the rigid shell itself, with millimeter accuracy:

• Volunteer 1: mean distance from MRI markers to the lower anterior corner of the S4 vertebra was 210.7 mm, with a standard deviation of just 1.8 mm
• Volunteer 2: mean distance to the S2 vertebra was 196.4 mm, with a standard deviation of just 1.4 mm

The Crucial Comparison: With Device vs. Without Device

To demonstrate the added value of their system, the researchers performed a comparison experiment. Since there is no standard repositioning system for MRI in routine clinical practice, they mimicked the radiotherapy tattoo technique: two landmarks were placed on each volunteer's body (on the right and left hip) and two landmarks on the MRI table. The volunteer then positioned himself, and two operators helped him align the body landmarks with the table landmarks. The skin-to-table alignment was visually confirmed to better than 1 mm precision.

The results were striking. Without the immobilization device, the median 3D displacement vector was 4 to 6 times larger than with the device:

• Volunteer 1 without device: median 3D displacement of 9.8 mm (range 4.6 to 10.5 mm), compared to 1.5 mm with the device
• Volunteer 2 without device: median 3D displacement of 8.5 mm (range 6.8 to 14.5 mm), compared to 2.3 mm with the device

The standard deviation of the distance between the MRI table marker and the sacral vertebra confirmed the same story. Without the device, the standard deviations were 9.5 mm (volunteer 1) and 6.6 mm (volunteer 2) — versus 1.8 mm and 1.4 mm with the device. In other words, the mold reduced day-to-day positioning variability by roughly 5 to 7 times. The researchers noted that, as expected, treatment without such a device would simply not be accurate enough for this purpose.

Ultrasound Targeting: Can the Device Aim the Heat Correctly?

The device needed to do more than hold the patient still — it had to hold an ultrasound applicator so that a future therapeutic beam would hit the same tissue region every time. The researchers tested this by attaching the conventional ultrasound imaging probe into the dedicated holder on the 3D-printed mold (an elastic belt wrapped around the two-part device for the session) and comparing new images to the reference images taken before the mold was created.

The agreement in target repositioning was high. The right iliac vessels appeared in the same anatomic plane, visualized at similar depth within 2 mm precision — as illustrated across 14 image panels in the published paper (Figure 6A–G for the first volunteer, H–N for the second). The authors describe this as a measure of how a future ultrasound hyperthermia applicator could appropriately target the region that needs to be heated.

Finally, the device proved fast and user-friendly in practice. The time needed for both volunteers to self-align in the immobilization device and get comfortable was always less than 1 minute across all 28 measurement sequences.

Clinical Implications: What This Means for Patients

This study provides encouraging early evidence that a single, personalized 3D-printed device could serve both radiation therapy and MR-guided focused ultrasound hyperthermia in prostate cancer patients. Specifically:

• Sub-millimeter precision is achievable without tattoos or lasers. The mold alone guided the patient into place, eliminating the need for permanent skin marks.
• The biggest source of error is along the body's length. The longitudinal axis (head-to-feet) showed the largest deviations. This matters because some machines, such as MR-LINACs with fixed flat couches, cannot perform rotational corrections.
• Even small rotations matter. The authors note that uncorrected rotational errors can significantly affect organ sparing — for example, the rectal V50 dose parameter (the volume of rectum receiving 50 Gy or more) in prostate cancer patients. A rigid mold that limits rotation could therefore help protect the rectum and other healthy tissues.
• Both thin and heavier patients can be accommodated. The system worked for volunteers with BMIs of 22 and 31. However, the weight loss experienced by volunteer 2 between scanning and measurement (BMI 31.4 to 29.8) increased rotational errors — a reminder that significant weight changes during treatment could affect how well the mold fits.
• Ultrasound aiming appears reliable. The ability to image the same vessel within 2 mm through the embedded holder suggests that therapeutic ultrasound beams could be aimed consistently across sessions.

For patients, this research points toward a future in which a personalized plastic mold — created from a quick surface scan — could streamline combination treatments. Instead of multiple separate positioning systems and skin tattoos, one device could be used for radiation planning, daily radiation sessions, and MRI-guided hyperthermia sessions, potentially making the entire course of treatment faster, more comfortable, and more precise.

Limitations: What This Study Could Not Prove

Readers should understand the important limitations of this study:

• Only two healthy volunteers participated. This is a proof-of-concept study, not a clinical trial. It involved volunteers without prostate cancer, so it cannot directly show benefits or side effects in actual patients.
• No actual hyperthermia was delivered. The ultrasound probe used was an imaging probe, not a therapeutic heating applicator. The study measured positioning and targeting accuracy, not treatment effectiveness.
• Different measurement systems gave somewhat different numbers. For volunteer 1, the median 3D error was 4.5 mm on OSM versus 1.5 mm on MRI; for volunteer 2, it was 0.8 mm versus 2.3 mm. OSM measures the external body surface, while MRI co-registration measures internal bony anatomy — and the two can differ, for example due to soft tissue changes.
• Weight changes affect fit. One volunteer lost weight during the study (BMI from 31.4 to 29.8), which increased rotational errors. Patients who gain or lose weight during a multi-week treatment course might need a new mold.
• Production is time-consuming. Printing each device took 65 to 77 hours, which could be a logistical barrier in routine clinical practice, though this is a one-time cost per patient.
• The prototype tested the device on a conventional MRI and LINAC, not yet in a combined MR-LINAC environment with real patients undergoing both treatments simultaneously.

Recommendations: What Patients Should Know

Based on this research, patients undergoing (or considering) radiation therapy for prostate cancer may want to keep the following points in mind:

• Precision is central to success. Daily, reproducible positioning during radiation therapy is associated with better long-term disease control and fewer radiation-induced side effects. Ask your radiation oncology team how your position is verified each day — with tattoos, surface imaging, X-ray imaging, or a combination.
• Ask about tattoo-free options. Many modern clinics use surface-guided radiation therapy (like the AlignRT system used here) that can reduce or eliminate the need for permanent skin tattoos. If tattoos are a concern for you, discuss alternatives with your care team.
• Mention any planned weight changes. Because custom immobilization devices fit the body snugly, significant weight loss or gain during treatment can reduce their accuracy. If you plan to change your weight during treatment, tell your radiation oncologists.
• Hyperthermia is an emerging adjunct, not yet standard for this setting. Combining heat with radiation is clinically accepted for some cancers and is being actively studied for prostate cancer. This specific device is a research prototype, not yet available in clinics.
• Know your numbers. Roughly 50% (1 in 2) of patients who receive salvage radiation after prostatectomy experience disease progression within 5 years, and toxicity risks rise above 70 Gy. If you are in this situation, ask your doctor whether hyperthermia or other dose-modifying approaches might be appropriate for you — and whether any clinical trials are available.

Frequently Asked Questions

What is a personalized 3D-printed immobilization device for prostate cancer radiation and heat therapy?

It is a custom rigid body mold created from a 3D surface scan of your pelvis. In early testing on healthy volunteers, it held the body still within about half a millimeter, reduced positioning errors by up to six times compared with tattoo-and-laser methods, and included a holder for an ultrasound applicator. It is still a research prototype.

How would this device help if I need radiation therapy after prostate removal?

After prostate surgery, a PSA rise can indicate remaining cancer cells, and radiation is a standard treatment. Precise daily positioning is crucial because better targeting may improve disease control and reduce bladder or rectal side effects. This mold aims to lock the body in the same position for both radiation and experimental MRI-guided heat therapy sessions, without needing tattoos or lasers.

How accurate was the 3D-printed mold in positioning patients during the tests?

In tests on two healthy volunteers, most positioning errors were below one millimeter in vertical and lateral directions. Overall 3D displacement on the MRI was a median of 1.5 to 2.3 millimeters. The largest errors were along the head-to-feet direction, which ranged up to about 8.5 millimeters in some repeated scans.

Did the 3D-printed device successfully aim an ultrasound probe at the same spot every time?

Yes, in the proof-of-concept study. When an ultrasound imaging probe was attached to the mold, the right iliac vessels appeared within 2 millimeters of their original reference location across many test sessions. This suggests a future therapeutic ultrasound beam could be aimed consistently, but no actual heat therapy was delivered in this experiment.

Is this 3D-printed mold available in clinics now for prostate cancer patients?

No, it is a research prototype tested on only two healthy volunteers, not prostate cancer patients. It has not yet been tested in a combined MR-LINAC environment or with actual hyperthermia delivery. More studies are needed before it becomes a routine clinical option. Printer time was also long—65 to 77 hours per mold—which could be a logistical barrier.

Should I get a second opinion about salvage radiation plus hyperthermia for prostate cancer after prostatectomy?

A second opinion can help you understand whether salvage radiation is right for you and whether hyperthermia—which uses MRI-guided focused ultrasound to heat tissue to 41–43°C and make radiation more effective—is an appropriate option. Research shows about half of patients receiving salvage radiation after prostatectomy experience disease progression within 5 years, and radiation above 70 Gy may increase side effects. Hyperthermia is being studied for this situation, and the device described remains a research prototype. Ask whether hyperthermia or clinical trials might fit your case. Diagnostic Detectives Network provides independent expert second opinions.

Source Information

Original article title: Magnetic resonance-guided ultrasound hyperthermia for prostate cancer radiotherapy - 2022

Authors: Pauline C Guillemin, Giovanna Dipasquale, Johannes WE Uiterwijk, Maud Jaccard, Orane Lorton, Pelagia Tsoutsou, Joanna Gariani, Pierre-Alexandre Poletti, Rares Salomir, and Thomas Zilli (co-first authors: Guillemin and Dipasquale)

Affiliations: Faculty of Medicine, Geneva University; Division of Radiation Oncology, Geneva University Hospital; Department of Radiology & Medical Informatics, Geneva University Hospital, Geneva, Switzerland

Journal: Journal of 3D Printing in Medicine (2022), Volume 6, Issue 2, pages 55–67

Publication details: First draft submitted August 4, 2021; accepted for publication February 15, 2022; published online March 8, 2022. DOI: 10.2217/3dp-2021-0024

Funding/Access: Published under a Creative Commons Attribution-NonCommercial-NoDerivatives (CC BY-NC-ND 4.0) license.

This patient-friendly article is based on peer-reviewed research. It is intended for educational purposes and does not constitute medical advice. Patients should discuss their individual treatment options with their oncology care team.